A2 Preparation Guide (Final)
Exam Duration: 3 hours
Total Marks: 180
I. General Preparation Principles
- Do all tutorial questions until you can perform them without notes.
- Revise all class examples.
- Study the theory supporting each practical question.
- Use the Survival Kits at the end of chapters.
- Check the Table of Contents in the eBook and ensure topic recognition.
- Do not “spot” topics. Everything is examinable unless stated otherwise.
- Treat old notes cautiously. Some legacy model answers are wrong or outdated.
II. What You Do Not Need to Memorize
| Item | A2 | A3 |
|---|---|---|
| Statistical tables | Provided | Provided |
| Formula sheet (as in A1) | Provided | Provided |
| Excluded | Excluded | |
| SimTalk syntax | Excluded | Excluded |
| TPS object definitions | Excluded | Excluded |
| Monte Carlo & Contaminating System (NTS) | Excluded | Excluded |
| Kolmogorov–Smirnov (K–S) test | Excluded | Included |
| Procedure Kim–Nelson (KN) | Properties only | Properties only |
III. Core Study Areas
1. Theory — Q1 (33 Marks)
- Keywords: Define, Explain, List, Why?, How did you … ?
- Covers simulation fundamentals, conceptual reasoning, and model understanding.
- Expect questions linking theory to model logic (e.g., why throughput behaves a certain way).
2. Input Data Analysis: χ², Inverse Transform — Q2 (36 Marks)
Focus on:
- Chi-squared test (χ²) for testing distributional fit.
- Inverse transform sampling for generating random variables (continuous and discrete).
Supporting notes (context only, not necessarily direct questions):
- ODF — discrete variable based on failure counts.
- Decision variables and space: p.97
= specific decision variable combination. (a single solutions combination of decision variables)
= decision space (all possible ). (set of all feasible combinations of decision variables)
- ξ — stochastic components such as service or failure times.
Excluded: Monte Carlo, K-S test.
3. Models — Q3–Q9 (103 Marks)
Know Model 0 (Drive-through / McD) and all other models.
| Concept | Focus |
|---|---|
| Throughput limits | All models capped by source (tap) rate. |
| Buffer Allocation Problem (BAP) | Behaviour as buffers → ∞. |
| Trauma Unit Model (TUM) | Identify entities and apply Shannon’s world view. |
| Events | Instantaneous occurrences (arrive, start, finish, leave). |
| System State Variables | Throughput, WIP, utilization over time. |
| Validation inside models | Recognise throughput effects when configuration changes. |
4. Validation and Conceptual Questions
Goal: Adequate representation, not “accurate.”
Discuss real tests, not theory lists.
Reference notes: Validation
| Model | Validation Focus |
|---|---|
| TUP (Trauma Unit Problem) | “Stay here after job” invalidates model logic — must leave to scrub/rest. |
| (r, Q) Inventory Model | Low r and Q can still cause high total cost (frequent ordering). |
| Buffer sizing | Observe upper/lower throughput bounds when changing buffers. |
5. Output Analysis — Q10 (8 Marks)
- Half-width (
) and required sample size ( ) interpretation.
- Multi-Objective Optimization (MOO): identify Pareto-optimal trade-offs.
- p-Table: p.78
- A pairwise comparison table, not ANOVA.
- Interpret p-values:
p > 0.05→ no significant difference.
p ≤ 0.05→ significant difference.
- Used to identify statistically similar or superior systems.
- A pairwise comparison table, not ANOVA.
- Experiment Manager (EM): outputs and interpretation are examinable.
6. Algorithms and Procedures
| Procedure | Required Knowledge |
|---|---|
| Genetic Algorithm (GA) | Crossover / mutation / population size ↔ exploration; generations ↔ exploitation. |
| Procedure Kim–Nelson (KN)p.81 | A procedure performing just enough replications to separate systems statistically — contrasts with ANOVA’s fixed replications. Know properties, not steps. |
IV. Paper Layout
| Q | Content | Marks | Focus |
|---|---|---|---|
| 1 | Theory | 33 | Define, Explain, List, Why?, How … |
| 2 | Input Data Analysis | 36 | χ², Inverse Transform |
| 3–9 | Models | 103 | Model 0, TUM, BAP, (r,Q), Events, Validation |
| 10 | Output Analysis | 8 | h, n*, MOO, p-table interpretation |
Final Advice
- Understand why systems behave as they do.
- Link observations to throughput, variability, and adequacy of representation.
- Apply reasoning, not recall.